Disease identification and measurement system and method

Abstract

The invention discloses a disease identification and measurement system and method, and relates to the technical field of road and tunnel disease identification and measurement. The system comprises an image acquisition module, a data temporary storage module and a disease identification and measurement module, the image acquisition module is arranged over against a to-be-detected object and is used for acquiring image data of the to-be-detected object; the data temporary storage module is used for temporarily storing the acquired image data; a YoV8 target detection model, an image processing algorithm and a camera imaging model are loaded in the disease identification and measurement module; the YoloV8 target detection model is used for identifying disease information in image data, the image processing algorithm is used for optimizing positions of diseases in images, and the camera imaging model is used for calculating size information of each disease; the invention further provides a corresponding method. A visual mode is adopted, equipment arrangement is convenient, and the detection efficiency and the recognition rate are high.

Description

Technical Field

[0001] The present invention relates to the technical field of road and tunnel disease identification and measurement, and in particular to a disease identification and measurement system and method. Background Art

[0002] The statements in this section merely provide background information related to the present disclosure and may not constitute prior art.

[0003] With the rapid development of my country's economy, the construction scale of transportation infrastructure has been continuously expanded. As an important part of transportation, roads and tunnels are of vital importance for their safe operation. However, due to the long-term influence of vehicle loads, environmental factors, material aging and other factors, various diseases will inevitably occur in road and tunnel structures, such as Figure 4 As shown, these diseases include: cracks, falling blocks and rust. If these diseases are not detected and treated in time, they will seriously affect the bearing capacity and service life of roads and tunnels, and may even cause major accidents such as road damage, collapse, and tunnel collapse.

[0004] Traditional methods for detecting road and bridge defects are often done manually, including visual inspection and wire-pulling methods. The visual inspection method mainly involves inspectors observing the images of various road and bridge components with the naked eye or using simple tools (such as magnifying glasses, flashlights, etc.) to conduct detailed inspections of various parts. The wire-pulling method is that inspectors use thin wires to straighten the two sides of the cracks in the areas where cracks may exist, and then use a ruler to measure the length and width of the cracks. Traditional defect detection methods can detect defects to a certain extent, but they generally have problems such as low efficiency, low accuracy, and certain damage to the structure.

[0005] In addition to traditional detection methods, the current popular method is to use machine vision to detect and identify diseases, which mainly includes traditional image processing technology and deep learning technology. Traditional image processing technology mainly includes feature extraction, edge detection, image segmentation, etc. This method has a fast processing speed, but the recognition accuracy is low. Deep learning technology mainly includes Mask R-CNN, YoloV8, SAM2 and other image segmentation methods. Although these methods have high detection accuracy, the segmented area is generally large, and the accuracy of crack detection and measurement is insufficient. Summary of the invention

[0006] The purpose of the present invention is to provide a disease identification and measurement system and method for the problems existing in the prior art, which adopts a combination of YoloV8 and traditional image processing to identify and measure the diseases, thereby solving the above problems.

[0007] The technical solution of the present invention is as follows:

[0008] A disease identification and measurement system, comprising: an image acquisition module, a data temporary storage module and a disease identification and measurement module;

[0009] The image acquisition module is arranged on the object to be detected and is used to acquire image data of the object to be detected;

[0010] The data temporary storage module is used to temporarily store the collected image data; the disease identification and measurement module is loaded with the YoloV8 target detection model, image processing algorithm and camera imaging model; the YoloV8 target detection model is used to identify disease information in the image data, the image processing algorithm is used to optimize the position of the disease in the image, and the camera imaging model is used to calculate the size information of each disease.

[0011] Furthermore, the image acquisition module includes: a camera, a lens, and a device for carrying and moving the camera.

[0012] Furthermore, the camera and the lens are mounted on a camera bracket, and the camera bracket is placed on the bottom or side to be detected.

[0013] Furthermore, the data temporary storage module includes: an industrial computer; the industrial computer is used for camera control and temporary storage of image data.

[0014] Furthermore, the disease identification and measurement module is deployed on a server.

[0015] Furthermore, the server and the industrial computer are both provided with WIFI modules, and data transmission is performed through the WIFI modules.

[0016] The present invention also proposes a disease identification and measurement method, based on the above-mentioned disease identification and measurement system, comprising:

[0017] Step S1: Acquire image data of the object to be measured through a camera;

[0018] Step S2: Identify diseases based on the YoloV8 target detection model;

[0019] Step S3: Optimizing the location of the disease in the image based on an image processing algorithm;

[0020] Step S4: Calculate the size information of each disease based on the camera imaging model.

[0021] Furthermore, the camera is calibrated before shooting to calibrate the internal parameters of the camera;

[0022] The training steps of the YoloV8 target detection model are as follows:

[0023] Step A: Training data collection: Use a camera to collect a certain amount of disease image data as a training data set for the YoloV8 target detection model;

[0024] Step B: Dataset annotation: annotate the damage information in the training dataset image and save the corresponding annotation data file; the damage information includes: cracks, chipping and rust;

[0025] Step C: Model training; load the pre-trained model that comes with YoloV8, use the labeled data file to train the pre-trained model, and after the training is completed, obtain the YoloV8 target detection model that can be used for disease identification.

[0026] Furthermore, the step S3 comprises:

[0027] Step S31: extracting the damaged image area; generating an image data containing only the damaged information according to the damaged location identified by the YoloV8 target detection model;

[0028] Step S32: image contrast enhancement: using gamma correction to enhance the image contrast of the image generated in step S31;

[0029] Step S33: using the OTSU algorithm to segment the contrast-enhanced image into a binary disease image containing only the disease;

[0030] Step S34: disease contour detection; perform connected domain analysis on the binary disease image and calculate the pixel area of ​​each connected domain; then filter the connected domain and set the pixel value of the filtered area to 0; finally, calculate the edge information of each connected domain based on the Canny edge detection algorithm, and sort the edge points counterclockwise to form a certain order of disease contour points.

[0031] Furthermore, the step S4 comprises:

[0032] Step S41: Calculate the three-dimensional space coordinates of the defect contour point in the camera coordinate system according to the internal parameters of the camera and the distance of the specific shooting surface of the camera;

[0033] Step S42: Calculate the size information of each disease based on the three-dimensional space coordinates;

[0034] The step S42 comprises:

[0035] For the two diseases of chipping and rust, the area information is counted; according to the corresponding binary images, the three-dimensional spatial coordinates of each disease contour point of the two diseases are calculated, and then the physical area value of each chipping or rust in the current disease image is obtained by using the triangle area integration method;

[0036] For crack diseases, the statistical information includes the length and maximum width of the crack; the calculation of the crack length first uses the three-dimensional spatial coordinates corresponding to the disease contour points of the crack to calculate the physical distance between every two disease contour points, and then calculates the length of the entire contour based on the integration principle. Finally, the length is divided by 2 to obtain the physical length information of the current crack; the calculation method of the maximum width of the crack is: first, obtain the corresponding three-dimensional spatial coordinates of the disease contour points of each crack, then traverse the disease contour points of each crack, calculate the disease contour point with the minimum distance from the disease contour point, and save the distance value until the statistical position of all contour points is completed, and finally, according to the statistical value, find a maximum distance value, which is the maximum width value of the crack.

[0037] Compared with the prior art, the present invention has the following beneficial effects:

[0038] 1. The present invention is superior to the detection method described in the background art in terms of measurement accuracy and efficiency, and can predict the subsequent changing trends of various diseases through the database, which is convenient for inspection and maintenance.

[0039] 2. The prior art uses manual measurement, which can detect diseases to a certain extent, but has low measurement efficiency, low accuracy, and may damage the object being measured; the present invention uses a visual method, which is convenient for equipment layout and has high detection efficiency and recognition rate.

[0040] 3. The existing technology uses deep learning to identify diseases. Although it can achieve a high recognition rate, the measurement accuracy of the disease is low. The present invention combines YoloV8 and traditional image processing methods, which can not only achieve a high recognition rate, but also accurately locate the position of each disease in the image, and calculate the specific size information of the disease through calibrated camera parameters, providing a basis for the prediction of the subsequent development trend of the disease. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1 This is a schematic diagram of a disease identification and measurement system;

[0042] Figure 2 It is a flow chart of a disease identification and measurement method;

[0043] Figure 3 This is a comparison of YoloV8 crack segmentation. The left one is the original image and the right one is the effect image.

[0044] Figure 4 Schematic diagram of common diseases. DETAILED DESCRIPTION

[0045] It should be noted that relational terms such as "first" and "second" are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0046] The features and performance of the present invention are further described in detail below in conjunction with the embodiments.

[0047] Embodiment 1

[0048] See also Figure 1 , a disease identification and measurement system, comprising: an image acquisition module, a data temporary storage module and a disease identification and measurement module;

[0049] The image acquisition module is arranged on the object to be detected and is used to acquire image data of the object to be detected;

[0050] The data temporary storage module is used to temporarily store the collected image data; the disease identification and measurement module is loaded with the YoloV8 target detection model, image processing algorithm and camera imaging model; the YoloV8 target detection model is used to identify disease information in the image data, the image processing algorithm is used to optimize the position of the disease in the image, and the camera imaging model is used to calculate the size information of each disease.

[0051] In this embodiment, specifically, the image acquisition module includes: a camera, a lens, and a device for carrying and moving the camera; the camera and the lens are installed on a camera bracket, and the camera bracket is set facing the object to be detected.

[0052] In this embodiment, specifically, the data temporary storage module includes: an industrial computer; the industrial computer is used for camera control and image data temporary storage.

[0053] In this embodiment, specifically, the disease identification and measurement module is deployed on a server.

[0054] In this embodiment, specifically, the server and the industrial computer are both provided with WIFI modules, and data transmission is performed through the WIFI modules.

[0055] See also Figure 2The present invention also proposes a disease identification and measurement method, based on the above-mentioned disease identification and measurement system, which specifically includes the following steps:

[0056] Step S1: Acquire detected image data through a camera;

[0057] Step S2: Identify diseases based on the YoloV8 target detection model;

[0058] Step S3: Optimizing the location of the disease in the image based on an image processing algorithm;

[0059] Step S4: Calculate the size information of each disease based on the camera imaging model.

[0060] In this embodiment, specifically, the camera is calibrated before shooting to calibrate the internal parameters of the camera;

[0061] It should be noted that the internal parameters of the camera can be expressed by K, as shown in the following formula:

[0062]

[0063] where f x , f y is the scaling factor, u 0 and v 0 Indicates the coordinates of the center of the image pixel, the unit is pixel, s is the non-vertical factor, which is generally not considered. Camera calibration is to solve the specific values ​​of each unknown parameter in the matrix K. The specific process is as follows:

[0064] (1) Prepare the calibration plate: Select a black and white checkerboard calibration plate with a specific size and ensure that the calibration plate has good flatness to ensure calibration accuracy;

[0065] (2) Take images: Place the calibration plate at different positions, angles, and distances, and use the camera to be calibrated to take several images. Ensure that the corners of the chessboard in the image are clearly visible and cover the entire image.

[0066] (3) Corner detection: Perform corner detection on the captured image and use the Shi-Tomasi corner extraction algorithm to extract the coordinates of the corners of the chessboard;

[0067] (4) Sub-pixel refinement: Use the LM (Levenberg-Marquardt) optimization algorithm to perform sub-pixel refinement on the detected corner points to improve the accuracy of the corner point coordinates;

[0068] (5) Solving the camera's internal parameter matrix K: Use the coordinates of the checkerboard corner points extracted from the image and their corresponding three-dimensional spatial coordinates to construct the solution equation for the K matrix. When the number of images from which the checkerboard corner points are successfully extracted is greater than or equal to 3, the camera's internal parameter matrix can be calculated using Zhang Zhengyou's plane calibration method;

[0069] (6) Solution of camera distortion parameters: In addition to internal parameters, cameras generally have distortion parameters. The distortion parameters of the camera are obtained by minimizing the reprojection error, as shown in the following formula:

[0070]

[0071] Where n represents the number of images taken, and i = {1, 2,,,n}, m represents the number of corner points detected in the i-th image, and j = {1, 2,,,m}, ||·|| 2 is the Euclidean distance between two points. m ij represents the actual pixel coordinates of the jth corner point in the i-th image, represents the pixel coordinates of the corner points calculated after projection transformation. K represents the camera intrinsic parameter matrix, k 1 ,k 2 ,p 1 ,p 2 are the distortion parameters of the camera, both of which are constants. i is the rotation matrix from the world coordinate system to the camera coordinate system in the i-th image, t i is the translation vector from the world coordinate system to the camera coordinate system in the i-th image. Through this formula, the Euclidean distance between the actual point and the projected point is calculated and minimized to obtain the distortion parameters and optimize the previously calibrated camera internal parameters.

[0072] In this embodiment, specifically, the training steps of the YoloV8 target detection model are as follows:

[0073] Step A: Training data collection: Use a camera to collect a certain amount of disease image data as a training data set for the YoloV8 target detection model;

[0074] Step B: Dataset annotation: annotate the damage information in the training dataset image and save the corresponding annotation data file; the damage information includes: cracks, chipping and rust;

[0075] Step C: Model training; load the pre-trained model that comes with YoloV8, use the labeled data file to train the pre-trained model, and after the training is completed, obtain the YoloV8 target detection model that can be used for disease identification; It should be noted that the disease identification process mainly includes: feature extraction, feature fusion, category prediction and non-maximum suppression, and disease information output;

[0076] Furthermore, for erroneous image data or unrecognized image data in the process of disease information recognition, they are used as input to the YoloV8 target detection model, and the trained YoloV8 target detection model is adjusted to meet the image recognition requirements under the basic environment, that is, the model is fine-tuned.

[0077] In this embodiment, it should be noted that the YoloV8 target detection model can only roughly segment the location of the defect in the image. For example, cracks Figure 3 As shown in the figure, the crack width identified by the YoloV8 target detection model is obviously wider than the actual crack width, which is not conducive to the subsequent determination of the danger of the crack. Therefore, the above recognition results need to be further optimized;

[0078] Specifically, the step S3 includes:

[0079] Step S31: Extract the defect image area; generate an image data containing only the defect information according to the defect location identified by the YoloV8 target detection model; the specific method is: for each type of defect, predefine a pure black image of the same size as the original (hereinafter referred to as the "predefined image"). Traverse each pixel in the original image, if the pixel is in the defect area, copy the value of the pixel to the new predefined image until all pixels are traversed;

[0080] Step S32: image contrast enhancement; in order to more accurately identify the location of the disease in the predefined image, the image generated in step S31 is enhanced in image contrast by using a gamma correction method; it should be noted that the mapping formula of the gamma correction is:

[0081] f(x,y)=I(x,y)^γ

[0082] In the above formula, I(x,y) represents the pixel value at the coordinate (x,y) in the original image, γ represents the coefficient constant in the Gamma correction, and f(x,y) represents the image pixel value at the corresponding position after correction. In order to enhance the contrast of the image, γ=2 can be selected in actual use;

[0083] Step S33: Use the OTSU algorithm to segment the contrast-enhanced image into a binary disease image containing only the disease. Different from the traditional OTSU algorithm, when calculating the maximum inter-class variance σ, the background with a pixel value of 0 needs to be ignored. The specific implementation steps are as follows:

[0084] (1) First, count the maximum pixel value I of the image max and the minimum pixel value I min, and the total number of pixels between them is N, with an average value of m G ;

[0085] (2) Assume that the segmentation threshold is T(I min ≤T≤I max ), where the number of pixels smaller than the segmentation threshold T is N 1 (target), that is, the probability that this part belongs to the target is: P 1 =N 1 / N. The number of pixels greater than the segmentation threshold T is N 2 (background), that is, the probability that this part belongs to the background is: P 2 =N 2 / N. Define the pixel mean of target and background as m 1 and m 2 ;

[0086] (3) According to the definition of between-class variance: σ 2 =P 1 (m 1 -m G ) 2 +P 2 (m 2 -m G ) 2 The larger the inter-class variance between the background and the target, the greater the difference between the two parts of the image. Therefore, for each threshold T, an inter-class variance σ is calculated. The threshold T corresponding to the maximum inter-class variance σ is the required segmentation threshold;

[0087] (4) Use the best segmentation threshold obtained in the previous step to segment the image into a binary image. Specifically, if a pixel value is greater than the threshold, then the pixel value is set to 0, otherwise the pixel value is set to 255. At this point, the binary image data containing only the disease obtained by the OTSU algorithm segmentation can be obtained;

[0088] Step S34: Disease contour detection; perform connected domain analysis on the binary disease image and calculate the pixel area size of each connected domain; then filter the connected domain and set the pixel value of the filtered area to 0; finally, calculate the edge information of each connected domain according to the Canny edge detection algorithm, sort the edge points counterclockwise to form a certain order of disease contour points; that is, first use the "two-pass scanning method" to perform connected domain analysis on the binary disease image and calculate the pixel area size of each connected domain. Then filter the connected domain, and according to previous experimental experience, for connected domains with an area less than 20 or greater than 999999, they are considered to be interference areas, and the pixel value corresponding to the area is set to 0. Finally, calculate the edge information of each connected domain according to the Canny edge detection algorithm, sort the edge points counterclockwise to form a certain order of disease contour points, which is convenient for subsequent area calculations;

[0089] Through the above four steps, the specific location information of the identified disease in the image can be more accurately located, thereby improving the accuracy of subsequent disease measurement.

[0090] In this embodiment, specifically, step S4 includes:

[0091] Step S41: Calculate the three-dimensional spatial coordinates of the defect contour point in the camera coordinate system according to the internal parameters of the camera and the distance of the specific shooting surface of the camera; that is, assuming that the distance of the specific shooting surface of the camera is D (the distance can be obtained by a laser rangefinder), and the image coordinates of a contour point of a defect in the image are (x, y), the three-dimensional spatial coordinates (X, Y, Z) of the contour point in the camera coordinate system can be calculated according to the parameters obtained by camera calibration, as shown in the following formula:

[0092]

[0093] where f x , f y is the scaling factor, u 0 and v 0 Indicates the center coordinates of image pixels, all in pixels. The specific values ​​of these parameters can be obtained during the camera calibration process;

[0094] Step S42: Calculate the size information of each disease based on the three-dimensional space coordinates;

[0095] The step S42 comprises:

[0096] For the two diseases of chipping and rust, only the statistical area information is needed; according to the corresponding binary images, the three-dimensional spatial coordinates of each disease contour point of the two diseases are calculated, and then the triangle area integration method is used to obtain the physical area value of each chipping or rust in the current disease image;

[0097] For crack diseases, the statistical information includes the length and maximum width of the crack; the calculation of the crack length first uses the three-dimensional spatial coordinates corresponding to the disease contour points of the crack to calculate the physical distance between every two disease contour points, and then calculates the length of the entire contour based on the integration principle. Finally, the length is divided by 2 to obtain the physical length information of the current crack; the calculation method of the maximum width of the crack is: first, obtain the corresponding three-dimensional spatial coordinates of the disease contour points of each crack, then traverse the disease contour points of each crack, calculate the disease contour point with the minimum distance from the disease contour point, and save the distance value until the statistical position of all contour points is completed, and finally, according to the statistical value, find a maximum distance value, which is the maximum width value of the crack.

[0098] In this embodiment, disease data statistics can also be performed; that is, the category information, image contour information, and corresponding measurement values ​​and original image data of each disease are stored in the database. In the database, the disease data detected are sorted in chronological order, and the measurement data of the same disease at the same location are compared to see if they have changed. According to the change values ​​of the measurement data of the same disease at the same location at different times, the subsequent change trend of the disease is predicted, which provides a theoretical basis for the subsequent grade determination of the disease and whether the disease needs to be treated.

[0099] It should be noted that all road, bridge and tunnel defects that are identified and measured using the above-mentioned system and method are included in the protection scope of this application.

[0100] The above-mentioned embodiments only express the specific implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the protection scope of the present application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the technical solution concept of the present application, and these all belong to the protection scope of the present application.

[0101] This background section is provided to generally present the context of the invention, and the work of the presently named inventors, the work to the extent described in this background section, and aspects of this section that did not constitute prior art at the time of application are neither explicitly nor implicitly admitted to be prior art to the present invention.

Claims

1. A disease identification and measurement system, characterized in that: include: Image acquisition module, data temporary storage module and disease identification and measurement module; The image acquisition module is arranged on the object to be detected and is used to acquire image data of the object to be detected; The data temporary storage module is used to temporarily store the collected image data; the disease identification and measurement module is loaded with the YoloV8 target detection model, image processing algorithm and camera imaging model; the YoloV8 target detection model is used to identify disease information in the image data, the image processing algorithm is used to optimize the position of the disease in the image, and the camera imaging model is used to calculate the size information of each disease.

2. A disease identification and measurement system according to claim 1, characterized in that: The image acquisition module includes: a camera, a lens, and a device for carrying and moving the camera.

3. A disease identification and measurement system according to claim 2, characterized in that: The camera and the lens are mounted on a camera bracket, and the camera bracket is arranged facing the object to be detected.

4. A disease identification and measurement system according to claim 3, characterized in that: The data temporary storage module includes: an industrial computer; the industrial computer is used for camera control and temporary storage of image data.

5. A disease identification and measurement system according to claim 4, characterized in that: The disease identification and measurement module is deployed on a server.

6. A disease identification and measurement system according to claim 5, characterized in that: The server and the industrial computer are both provided with WIFI modules, and data transmission is performed through the WIFI modules.

7. A disease identification and measurement method, characterized in that: A disease identification and measurement system according to any one of claims 1 to 6, comprising: Step S1: Acquire image data of the detected object through a camera; Step S2: Identify diseases based on the YoloV8 target detection model; Step S3: Optimizing the location of the disease in the image based on an image processing algorithm; Step S4: Calculate the size information of each disease based on the camera imaging model.

8. A disease identification and measurement method according to claim 7, characterized in that: The camera is calibrated before shooting to calibrate the internal parameters of the camera; The training steps of the YoloV8 target detection model are as follows: Step A: Training data collection: Use a camera to collect a certain amount of disease image data as a training data set for the YoloV8 target detection model; Step B: Dataset annotation; Label the disease information in the training data set images and save the corresponding labeled data files; The damage information includes: cracks, chipping and rust; Step C: Model training; load the pre-trained model that comes with YoloV8, use the labeled data file to train the pre-trained model, and after the training is completed, obtain the YoloV8 target detection model that can be used for disease identification.

9. A disease identification and measurement method according to claim 8, characterized in that: The step S3 comprises: Step S31: extracting the damaged image area; generating an image data containing only the damaged information according to the damaged location identified by the YoloV8 target detection model; Step S32: image contrast enhancement: using gamma correction to enhance the image contrast of the image generated in step S31; Step S33: using the OTSU algorithm to segment the contrast-enhanced image into a binary disease image containing only the disease; Step S34: disease contour detection; perform connected domain analysis on the binary disease image and calculate the pixel area of ​​each connected domain; then filter the connected domain and set the pixel value of the filtered area to 0; finally, calculate the edge information of each connected domain based on the Canny edge detection algorithm, and sort the edge points counterclockwise to form a certain order of disease contour points.

10. A disease identification and measurement method according to claim 9, characterized in that: The step S4 comprises: Step S41: Calculate the three-dimensional space coordinates of the defect contour point in the camera coordinate system according to the internal parameters of the camera and the distance of the specific shooting surface of the camera; Step S42: Calculate the size information of each disease based on the three-dimensional space coordinates; The step S42 comprises: For the two diseases of chipping and rust, the area information is counted; according to the corresponding binary images, the three-dimensional spatial coordinates of each disease contour point of the two diseases are calculated, and then the physical area value of each chipping or rust in the current disease image is obtained by using the triangle area integration method; For crack diseases, the statistical information includes the length and maximum width of the crack; the calculation of the crack length first uses the three-dimensional spatial coordinates corresponding to the disease contour points of the crack to calculate the physical distance between every two disease contour points, and then calculates the length of the entire contour based on the integration principle. Finally, the length is divided by 2 to obtain the physical length information of the current crack; the calculation method of the maximum width of the crack is: first, obtain the corresponding three-dimensional spatial coordinates of the disease contour points of each crack, then traverse the disease contour points of each crack, calculate the disease contour point with the minimum distance from the disease contour point, and save the distance value until the statistical position of all contour points is completed, and finally, according to the statistical value, find a maximum distance value, which is the maximum width value of the crack.

Images